Diversity of Wild BeesSpecies Composition

Analyzing various sources that mentioned specimens labelled from Havana (Online resource1), we gathered 1,322 records, of which 1,067 are new records for the regions that had not been published. The highest number of individuals is in the collections of MNHNCu (46.1%), followed by IES (22.8%) and MNHN (16.9%). The KU-NHM collection represents 7% of the recorded individuals, and this data has been already published (https://www.amnh.org/research/invertebrate-zoology/collections/hymenoptera-apoidea). The other collections accounted for 7.2% of the recorded data.

In the three provinces, 52 species (52.6% of the Cuban wild bees described) of wild bees were recorded out of the 95 species documented for Cuba (Appendix 2, Online resource1). They belong to 23 genera (31 for Cuba) and all the 4 families present in the country. The number of species found in each historical collection was as follows: 46 species in MNHNCu, 27 species in IES, 26 species in both museums MNHN and KU-NHM, 6 species in AMNH, 3 species in USNM, 2 species in INHS, 2 species in MZC and 2 species in SI-NMNH, and 1 species in each place of CUIC, FSCA, LACM, NMNH, PCYU, USNM (Appendix 1, Online resource1).

Conservation Status of Wild Bee Species

Thirty-three species were common, i.e. with more than eight reports and seven species with more than 60 individuals: Lasioglossum parvum, Halictus. poeyi, Agapostemon femoralis, Agapostemon poeyi, Apis mellifera, Nomada cubensis, and Ceratina cyaniventris.

Some species were uncommonly recorded in these provinces. Only 16 species listed show less of eight reports i.e. Augochlora magnifica, Augochlora praeclara, Centris fasciata, Ceratina cockerelli, Coelioxys sannicolarensis, Coelioxys tridentatus, Colletes submarginatus, Epeolus pulchellus, Lasioglossum normalis, Megachile. apora, Meg. concinna, Meg. curta, Mesoplia cubensis, Triepeolus roni and T. wilsoni and Xeromelecta alayoi. Three species are classified as"Critically Endangered"on the Red List of Cuba: Coe. sannicolarensis, Coe. tridentatus and Hy. limbifrons.

Ecological Characteristics of Wild BeesOrigin and Distribution

The bee fauna found in this study consisted mostly of native species (90.4%, 47 species), with a smaller proportion of introduced species (9.6%, 5 species). The introduced species included A. mellifera, Mel. beecheii, Meg. concinna, Meg. lanata and Meg. rufipennis. Out of the total native species, 13 species (25%) were biogeographically defined as endemic (Appendix 2; Online resource1). These endemic species belong to the genera Anthophora (1), Ceratina (1), Epeolus (1), Melissodes (3), Mesoplia (1), Nomada (2), Hylaeus (1), Agapostemon (1), Lasioglossum (1) and Coelioxys (1) (Appendix 2, Online resource1). In addition to these Cuban endemics, 25 species are restricted to the Lesser Antilles (49%) and occur on various islands.

Distribution of Sampling in Western Cuba

Concerning collection localities of the 1018 records with complete collection locality data, 117 sampling localities were identified for 48 species of wild bees (4 species without complete locality data) in Havana, Mayabeque and Artemisa (Online resource1). In the province of Havana 51 localities were sampled, while Mayabeque comprises 38 localities, according to the current political-administrative division, and Artemisa was the one that sampled the least, with a total of 28 localities (Fig. 3).

Fig. 3

Distribution of wild bees in the three western Cuban provinces (occurrence map). Provinces: Havana (1), Artemisa (2), and Mayabeque (3)

The highest sampling frequency was observed in Santiago de Las Vegas in Boyeros and El Laguito in Playa (Havana), with 157 records encompassing 35 species. In Artemisa, El Mariel accounted for 99 out of 804 records and five species, whereas in Mayabeque, sampling was concentrated in Güines, yielding 64 records each for seven species (Fig. 4).

Fig. 4

Densely sampled site in Havana: El Laguito, Havana

Feeding Habits of Species and Seasonal Distribution of Sampling

The feeding habits of the species varied between polylectic non parasitic (80.8%, 42 spp.) and kleptoparasitic (19.2%, 11 spp) (See Appendix 2, Online resource1). The 11 kleptoparasitic species belong to the genera Epeolus pulchellus, N. cubensis, N. pilipes, N. cruralis, T. vicinus, T. roni, T. wilsoni, Coe. rufipes, Coe. sannicolarensis, Coe. tridentatus and X. alayoi.

When assessing the seasonal distribution of sampling, no significant differences (p-value = 0.9) were observed in the abundance of species per period (Fig. 5; Appendix 2). An analysis within months by season showed that May registered the highest number of records (156), during the rainy season, whereas September recorded the lowest (42). In the dry season, April recorded the highest number of collects (207), while November exhibited the lowest (18). The most collected species in each of these months were: L. parvum in April (41 ind), H. poeyi in May (21 ind) and Me. cubensis (4 ind) in November.

Fig. 5

Seasonal distribution of wild bee species occurrence. The color intensity gradient represents the distribution of catches for each species over the study period

The number of species per season was slightly higher in the rainy season (49) than in the dry season (39). Sixteen species of wild bees were present in only one season and 32 species in both seasons. The most abundant species were L. parvum (155 ind.), H. poeyi (121 ind.), and A. femoralis (81 ind.) in both periods. Six species were observed only in dry seasons and 10 only in rainy seasons.

Temporal Variation of Diversity, Dietary Habits and Phenology in Wild Bee Species Over 125 YearsDiversity Variation of Wild Bees. Persistence and Decline of Wild Bees Over Time

The oldest records trace back to the year 1867, featuring nine collected specimens belonging to three species (A. femoralis, A. elegans, and Ex. pulchella). The most recent species records for Havana are from 2024 including the species A. mellifera, C. cyaniventris, Ex. pulchella, H. poeyi, M. concinna, Mel. beechei and X. cubaecola. All data for 2024 were collected on urban farms as part of a project conducted by the first author.

Nineteen species were present in more than three periods. Eleven species were present in only one period, and seven species were present in all five periods: (A. femoralis, A. poeyi, A. mellifera, C. cyaniventris, Ex. pulchella, H. poeyi and N. cubensis). Of the species present in a single 25-year period, the 1976–2000 period was the most outstanding.

Throughout 125 years of data from different sources, the dominant (7), intermediate (25) and rare (20) species were determined according to the frequency of observation. The three dominant species were: L. parvum, H. poeyi, A. femoralis followed by A. poeyi, A. mellifera, N. cubensis, and C. cyaniventris. Among the rare species, 16 species had a frequency of observation lower than five times and four species were recorded only once (Coe. sannicolarensis (unknown), Ex. pulchellus (1975–2000), and Meg. apora (1975–2000).

When analyzing how the species dominance varied over the 125 years, it was observed that before 1900, A. femoralis (eight ind.) was the most collected bee, a species that appears among the most dominant in the wild bee population. In the 1900–1925 period, the most commonly collected species from the western region in zoological collections were A. mellifera and N. cubensis, followed modestly by C. aethiops. In the 25-year period that followed, the abundance of specimens declined, with Meg. rufipennis and A. poeyi standing out. In the 1951–1975 period, there was an explosion of L. parvum in the collections, followed closely behind by A. poeyi. In the following 25-year period, when more individuals were found, H. poeyi and L. mestrei were the most abundant species. In the most recent period, 2001–2025 the small bee, A. mellifera was the most collected, followed by C. cyaniventris.

In the analysis of the 52 species across the 125 years of data, it was observed that 31.5% of the species have not been observed since the last century, some for more than 100 years, such as C. cornuta (1909) and Aug. magnifica (1911). However, 68.1% of the species have been observed in recent years, even 7 species not previously collected were reported in this period in one of the provinces (Pinar del Río): Caupolicana nigrescens, C. tarsata, C. taina, Coe. productus, Hy. royesi, Meg. armaticeps and Meg. droegei.

Distribution of the Effort of Collects Over a Century

Species sampling, and our knowledge of their richness, relative abondance and ecology, have increased significantly over the past century, but have not yet reached a plateau. In the initial 25-year period 1900–1925, 26 species of wild bees were documented in Havana, Mayabeque and Artemisa provinces (Fig. 6). This number increased steadily throughout the century to the current count of 52 species. Sampling in the period 1951–1975 was remarkably extensive compared to the preceding 25-year period. No significant differences were found between the periods in terms of species richness and abundance, with the exception of the 25-years period, before 1900, which showed significant differences with the rest of the group (Table 1).

Fig. 6

Species accumulation curve of wild bees through four successive 25-year periods between 1900 and 1999

Table 1 Temporal variation in species richness and individual abundance of wild bees in western Cuba

Over the years, the collecting effort fluctuated (except in the second period) and the visited sites varied from one period to another. This effort was lowest in the period 1926–1950 and the highest number of collecting events was between 1951–1975. The successive periods exhibited the following number of collect events: 205 during 1900–1925, 61 during 1926–1950, 499 during 1951–1975, 272 during 1976–2000, and 156 during 2001–2025.

A species accumulation curve describing the richness of species through the century shows a cumulative increase in species richness as more years are sampled, with a final number of species of 52. The maximum expected number of species is estimated at 56 (± 4) with the Chao estimation (Fig. 7), and 60 (± 4) according to Jackknife. We recorded 92.9% and 86.7%, respectively, of the maximum predicted species count.

Fig. 7

Species accumulation curve combining captures from all study years. The blue line represents the asymptotic model on the species accumulation curve, and the black lines represent the Jackknife and Chao estimation of richness

Feeding Habit of Species Over the Years

Data from the four 25-year periods indicated a systematically higher presence of polylectic non parasitic species compared to kleptoparasitic species (Fig. 8). During 1900–1925, kleptoparasitic species reached their highest proportional abundance (22.7%), relative to polylectic species collected in the same period. However, this value decreased progressively in the following two periods, decreasing by 8% to 14.6% in 1926–1950 and even more to 9.77% in 1951–1975. A slight recovery was observed in 1976–2000, with an increase in kleptoparasitic species close to 15%. However, the downward trend resumed in the last period (2001–2025), when kleptoparasitic species reached the lowest proportional abundance ever recorded (3.85%).

Fig. 8

Proportional abundance of polylectic and kleptoparasitic wild bee species over five 25-year intervals in western Cuba

Seasonal Distribution of Species Over the Years

The distribution of sampling across the four 25-year periods (Fig. 9) reveals notable trends and imbalances in collection efforts over time. During the earliest period (1900–1925), nearly all records (99%) were collected during the rainy season, with only minimal sampling in the dry season. A similar, though slightly less pronounced, pattern was observed in the second period (1926–1950), where 72.1% of the collections occurred during the rainy season. In contrast, sampling efforts in the following periods (1951–1975, 1976–2000, and 2001–2025) were more evenly distributed across seasons, reflecting a shift toward capturing data in both rainy and dry conditions. For instance, in 1951–1975, 44% of collections were taken during the dry season, and in 1976–2000, dry season collections reached 50%. All species were recorded in both seasons.

Fig. 9

Seasonal distribution of the wild bee community throughout the surveyed time periods in the three western provinces of Cuba

In terms of seasonal patterns, the rainy season consistently showed higher sampling effort, particularly in the months of May and June. For example, in the period 1951–1975, 101 records were collected in June, while May accounted for 63 records. During the same period, the dry season saw its peak in April, with 126 records collected. Similar trends persisted in the later periods, though at lower magnitudes. Between 2001–2025, March (dry season) showed a notable increase in collections, with 54 records, while April and June (rainy season) contributed 43 and 11 records, respectively.